Further Isolation of the Silverstone PCIe - USB Card

[CoenP]

Mystery ground connection left aside, I do see some similarities between my setup and anthonys.

Both make use of a fully connected usb interface, that is including the usb ground.

The other is the 'single' ground. If I am correct, the pe acts as ground for Antonys setup (like a star earth) and in my setup the PCs pe is the only one connected (think bus earth). This reasoning is valid only if on the pc side the pci-e ground is connected to the chassis which is connected to the pe wire and the usb groundpoint in the nos1 is connected to the case (wich is connected to pe by 'the other' wire).

I think we are hearing a very similar improvement though through quite different methods.

Regards.

[PeterSt]

Regarding your question about the current for the FPGA when the ground wire was disconnected in the NOS1 and the loom was in place I am not sure that the USB card was visible to the operating system at that time because I did not check.

Anthony, no problem. In my case it's certainly visible.
Peter

[acg]

Well, I guess that I must have been doing something wrong yesterday when I tried to listen to music with my loom in place and the black earth wire removed in the NOS1 because today it works.  So I guess some of yesterday's theories are now redundant.

I don't know why I could not get it to work yesterday, but today it worked first time.

For clarity, I think that I should give a brief overview of the earthing arrangements of my system as it currently stands.

• The NOS1, XXHE PC, lab LPS and amps are all connected with three prong power cords into the same power socket which is a dedicated line for audio
• I have lifted the black wire from the NOS1
• The lab LPS is feeding 3.3V and 5V into the Silverstone USB card which has a special cable loom that attempts to isolate it completely from PC ground

Anthony

[acg]

And be careful with static!

regards, Coen

Well, you might be right here.  I have noticed a very intriguing phenomenon with the loom in place:  when there is music playing (NOS1 black wire either connected or disconnected) and I turn on or off one of the fluorescent lights in the room or even in the next room (through a brick wall) the music stops.  It seems as though the Silverstone usb card is no longer recognised by the computer and I have to stop XXHE from playing the music, turn off the computer totally and then turn it back on.

Normal light bulbs are no problem.  Remember that no lights are on the electrical circuit which powers the audio.

Unusual.  I have not experienced this before and will do a little research into fluoro lights.

EDIT:  The music does not go off when the fluoro light is flickering but it goes off when the light snaps on permanently.  When switching off the light the music stops straight away.

[acg]

An Update.

So I pulled out my cable loom and have experimented with the fluoro lights.  Regardless of whether I have use my lab LPS to power the Silverstone molex or not, I get pops through my speakers that sound similar to samples being lost.  These pops occur when the fluoro light flickers on and then when it strikes permanently and then when it shuts down. 

So the lab LPS does not seem to be part of the problem, but obviously my loom is causing a greater sensitivity to the fluoro light issue.

Anthony

[acg]

Solved.  I went into town and grabbed a digital ballast for one of the fluorescent lights to replace to magnetic ballast and starter.  Wired it all up, turned on the music flicked the switch...and...nothing...no pops...no music stopping...just how it should be.

But a nice lesson about things in the house that can impact your audio power supplies.

Anthony

[PeterSt]

Anthony,

Not solved I think. Better put back that "malfunctioning" device and with that as a measure find your lacking ground ...

IOW this is the *typicle* proof of lacking ground. Most probably your whole needed grounding comes from PE. Not good.

Make yourself such an extension block with x power outlets and cut the ground inside. Or better : cut it halfway when possible. So, the better ones have copper (bare) leads running along the PE points in the outlets, and if you cut that halfway then you can easily choose for PE or not (say 4 outlets now carry PE, 4 do not).

When not using PE for the PC now probably nothing works anymore (for sound). But maybe it does, because what your lights imply is changing the ground potential and that sudden *change* is not tolerated (comes down to e.g. 5V suddenly being 20V (or 0V) for a short period of time.
Remember, the more devices (not lamp bulbs I think) switched on, the more PE will lift. You can see this by means of two PE's (say two different earth pins), the one serving as a reference for the other.

Regards,
Peter

[acg]

Hi Peter,

You are most likely correct that the problem is not solved.  The 30 year old magnetic ballast that I replaced with a new digital ballast in one of the lights is known for injecting large voltage spikes into the house wiring when it turns on and turns off.  The digital ballast is apparently able to suppress these spikes.  I have always had these pops when turning the lights on and off but never investigated them before.

I will play around with the power-board idea and let you know what I find.

Regards,

Anthony

[acg]

So today I have played around with the earth arrangement of my audio gear.  When changing plugs I noticed how difficult the plugs were to get into the socket (very difficult) and even the switches were very difficult to move.  Also with one of the sockets, I could hear a buzzing noise when turning on a component so I was straight off into town and got myself a new power point and installed it before proceeding further.
The surprising result here was that my DC offsets in the NOS1 now settle closer to their optimum settings.

#	Silverstone Card	Black Wire in NOS1	XXHE PC	NOS1	LPS	AMPS	Effect of Lights
1	Isolated	Lifted	PE	PE	---	PE	Pops
2	Isolated	Lifted	PE	PE Lifted	---	PE	No Pops
3	Isolated	Lifted	PE Lifted	PE Lifted	---	PE	No Pops
4	Isolated	Lifted	PE Lifted	PE Lifted	PE	PE	No Pops
5	Isolated	Lifted	PE Lifted	PE Lifted	PE Lifted	PE	No Pops
6	Loom	Lifted	PE Lifted	PE Lifted	Lifted	PE	Computer won't start
7	Loom	Lifted	PE Lifted	PE Lifted	PE	PE	Computer starts but NOS1 DC offsets won't settle
8	Loom	Lifted	PE	PE Lifted	PE Lifted	PE	Pops then NOS1 loses DC offsets
9	Loom	Lifted	PE	PE Lifted	PE	PE	Pops then NOS1 loses DC offsets
10	Loom	In Place	PE	PE Lifted	PE	PE	Pops and eventually will lose lock with NOS1 which requires a restart but does not lose DC offsets
11	Loom	In Place	PE	PE Lifted	PE Lifted	PE	Pops and loses lock with NOS1 which requires a restart but does not lose DC offsets
12	Loom	In Place	PE Lifted	PE Lifted	PE Lifted	PE	Pops and will lose lock with NOS1 which requires a restart but does not lose DC offsets
13	Loom	In Place	PE	PE	PE	PE	Pops and will lose lock with NOS1 which requires a restart but does not lose DC offsets

Sound with the loom in is preferred to sound with the loom out.

Hope this helps.

Anthony

[CoenP]

Great work Anthony!

I think we may conclude that the Loom does not provide for enough grounding connection to make it work without some annoyance. This must have to to with an uncertain ground potential between the card and the motherboard.

Now what if we could actively FULLY isolate the PCIE card like with some fast opto couplers....hmmm...

I wondered: why didn't you experiment with the power amps PE disconnected? Than there will be no more (redundant) ground path via the power amps. Anyway this would mimic my setup (as an alternative to 2).

regards, Coen

[PeterSt]

Anthony, man, what have you all done !
Well, at least this proves that I myself have been working on a couple of things in the past and how the DC Offset metering got in; it forces you to have it correct (to some extend).

Now I must really digest all the combinations. So I will.

Peter

PS: Notice that Coen is correct to best avoid PE for the power amps. At least that is my experience over and over again (but it will not be a real rule).

[acg]

The DC offsets made it easy actually.  When they are not fixed and music is playing it is a quick dash to the amp power switch but after the first time that happened I actually turned off the amps when testing the lights and just watched the DC Offsets.  If it proved safe after a dozen switches of the lights or so I would turn the amps back on and listen for pops.  I learnt a lot in the process.

I will try the amps off PE at some stage but I thought that if I messed with them yesterday that I would have double the amount of testing!!!

Anyway, a few things came out of this I think:

• The loom does sound better.  This could be the dubious ground loops or the LPS or both
• The NOS1 should be run with lifted PE.
• If the loom is used the black wire in the NOS1 must be connected

Anthony

[PeterSt]

Anthony,

Remember this is with an isolated Silverstone card, which I did not like initially, but with further isolation from the PCIe bus the sound is sweet and delicate and dynamic and beautiful.  The "sssss's" sizzle and decay but don't bite.

I made a remark about this when you wrote this, but (sadly) I scratched it because I drifted off in that particular post (so the whole post wasn't posted). But my remark about this was :

The sizzle is a tough one to interpret;
For me it could be an indication (and I mean when I perceive that myself) that some higher frequency is riding on it all which make the S sizzle. Btw, same can happen with a nylon string although I'd call that "zzoooming";

In both cases (S and nylon string) this *is* a good characteristic and at least for the nylon string I will not be able to discern the better and the worse. To understand this, we better look at the sizzling S and try to see how much real that would be. I mean, try to produce that sizzle yourself; you will be able to do it, but it won't be easy (it's even hard to hear it yourself and another person needs to listen).
Now, when you have tried this and have seen how difficult it is, you can come to the conslusion that what you perceive from a random track with random singer with his/her random S that sizziling is not likely. Also, it needs "Z" to be incorporated which is why it hardly can be a natural S you're hearing.

This has been my perception for a longer time, but I must say that I never really "tuned" on it, because the sizzling seems to exhibit better resolution/resolvement and audibly it is not disturbing. Also, combine it with the nylon string which gets better from it (plus better discernable from the metal string) and you are fine. Still, when the chips are down like in these posts (about PE and such) I think it should be (?) a measure of the wrong. And of course it is placeboed in my mind because of no ground = no good.

It's only sad that I scratched this text at the time I wrote it which was way early in the topic because then I would have been predictive while now it seems in aftermath. But really not so.

Btw, I now wrote it after all bcause I tried to make a best choice of your table of happenings which I couldn't. So the best choice would be one which plops with your solution of not switching the lights but which right away also prooves that PE is used for signal ground. Notice that "signal ground" is of various types, only one of them real (music output) signal ground and you can see it by the DC Offset changing or not locking.

What I only start to recognize is how a PC itself can be the baddy in this all, because I can see that I myself can't get things wrong at this moment, while the whole (DC Offset) setup has been created with another PC which *did* have problems like not wanting to start. Not much useful for you all, but this same PC ended up elsewhere (not for music playback) and when it was powered down I had to ... detach USB devices (like a printer) before it ever wanted to boot.
I can also add the fairly unrelated (or not) happening that half a year or so ago, we were struck by direct lightning and everything survided expect that PC (main earth switch tripped). After that it never wanted to boot again and it was replaced.

Remember, it is only today that I am able to combine these things (or suggest they are related) and it can only happen from more "sightings" from guys like you, your experiments and your descriptions of it.
These things are mighty difficult and even while I'm focused on it, it apparantly takes years to nail it down. To some extend that is, but we won't stop experimenting.
Personally (but I don't see all) I think that the few forums posts in here - relative to say millions elsewhere - are so much more valuable than these millions of others.

Peter

[Arjan]

Hi Anthony,
Thanks for the loom description. I have the loom running now, but I prefer it with the (modified) Paul Pang card. Because with that card I like the (less) highs more and ALL details are there!
I have a second Silverstone and PP (in order) and will make a combination of the two.... See what that will bring.
And W7 is also really nice to hear again.....
Regards, Arjan

[AlainGr]

I suppose there must be at least one component connected to PE - so it would be the PC and PC only ?

Alain